A Dangerous New Investment Trend

Andrew Snyder
by Andrew Snyder, Editor-in-Chief, The Oxford Club
machine learning investment trend

Steve Jobs returned to Apple (Nasdaq: AAPL) in 1997. Seven years later, the nation was slimming down by cutting carbs. And in 2007, Netflix (Nasdaq: NFLX) began streaming video online.

These are all major events... major events that changed industries, moved share prices and made investors rich.

They are not predictable events that a complex algorithm or pattern-recognizing computer would have picked up. They are events that only humans who are paying attention to or are highly involved in the industry would see coming.

All of them prove a valuable point.

When we invest, we’re investing in companies with real humans at the helms, real products in the warehouses and real people as customers - all things computers aren’t so good at dealing with.

The topic is on my mind, as I’ve been studying a growing trend in the hedge fund industry. More and more funds are dumping millions of dollars - much of it is investor dollars - into machine learning.

Their goal is to build a machine that can predict where the economy will go next and then directly invest in the stocks it thinks will lead the charge... all without input from us oh-so-dull humans.

To be sure, computers have helped investors make boatloads of money. But after eight years of underperforming the market, Wall Street’s supposed best have taken it too far. They’re letting computers do all the work.

There is no human interpretation of results. No checks and balances. The machine does it all.

It’s dangerous.

We could all die, one manager said, and the computer would keep on trading.

To the techno-minded, it all sounds great. That a computer can learn from its mistakes and learn to decipher even the smallest patterns in the nearly infinite data feed the internet now makes available is quite a marvel.

It’s too bad, though, that the economy and the companies that act as its building blocks are run by emotional, ever-changing humans.

No computer knew that Apple’s board would suddenly put Jobs at the helm. No computer knew that Netflix was planning to stream video or that its unique programming would turn it into a major content creator. And no computer could predict that a single diet book would take the nation by storm, slamming the baking industry as we saw the Atkins diet do in 2004.

It takes humans... because human behavior molds the economy.

With machine learning, we’re watching an incredible fad unfold.

We fear it’s one that will cost investors large sums of money and introduce fresh volatility. Worse, we’re convinced it’s going to create a generation of programmers, not investors.

That’s dangerous.

After all, investing is as much about human psychology as it is numbers and equations. The numbers on a balance sheet mean nothing without the context of the business and the people they represent.

As investors in charge of our own financial fates, it’s up to us to understand the numbers - and to have the ability to decipher their values.

If not - if we simply allow a machine to do it for us - failure will be the final result. Computers will compete with each other until their trading on nothing remotely close to the fundamentals of the companies they’re investing in.

That’s dangerous. And it’s certainly not how I want my money working for me.

We must invest in real companies - with real humans in the corner office producing real products. Moreover, we must empower ourselves to have the skills necessary to understand our investments.

It’s the cornerstone of our philosophy.

Sure, we can use computers to guide us to the best opportunities. But we must not rely on them. It takes a human touch to understand a human economy.

Mark my words, computer learning and investing will be a nasty, volatile cocktail. The results won’t be good.

You can call me old-fashioned. But you won’t call me broke.

Good investing,


Thoughts on this article? Leave a comment below.

Lecture Notes